A file conversion/manipulation software for corpus linguistics.
- See the Github's wiki for documentation.
- Current version: 3.3.0 from 2025-02-14. For a complete list of changes, see the Changelog.
Corflow is a tool written in Python to (a) manipulate files or (b) change a file's format, mainly applying to files used in the context of corpus linguistics (oral linguistics) and multi-layered annotated corpora. It allows performing operations on a file's stored linguistic information from any supported format and to safe the changes as a file from any supported format.
As of today, Corflow supports the following file formats:
| Tool | Format |
|---|---|
| ELAN | .eaf |
| Praat | .TextGrid |
| EXMARaLDA | .xml |
| Pangloss | .xml |
| Transcriber | .trs |
Future releases are planned to include an import and export option from .csv files as well as from ANNIS.
Install Corflow via PyPI using pip by typing the following command into your terminal (within an active virtual environment):
pip install corflowTo learn how to use Corflow, visit the Github's wiki and take a look at the Corflow tutorial series provided by the AIRAL project.
If you encounter problems trying to install and use Corflow, please visit the first tutorial of the mentioned tutorial series.
- X-to-Y conversions: Conversions from any supported format X to another supported format Y, e.g. from ELAN's .eaf to Praat's .TextGrid, in the same manner as Pepper's Swiss Army knife approach.
- One underlying model: Manipulating a file's stored information from any supported format using the same underlying model.
- Lossless conversions: As little information as possible should be lost during converting a file.
- Accessibility: The package should be available for (a) automatic integration, (b) through command prompts and (c) a dedicated graphical interface.
- Even more accessibility: The package should require as few third-party libraries as possible, be easy to understand and to expand (by users adding their own scripts). The software's core audience is expected to have little to no experience with programming languages and writing code. More advanced users are expected to prefer Pepper.
Corflow, originally the multitool, has been started around 2015 to anonymize and convert files for the OFROM corpus (at Neuchatel, Switzerland). Initially in C++, it was reworked from 2016 to 2019 in the ANR-DFG SegCor project (at Orleans, France) and translated into Python. It was further developed from 2019 to 2022 within the ANR-DFG DoReCo project (at Lyon, France). At present, it is actively developed and used for the DoReCo corpus within the AIRAL project (at ZAS Berlin).
- No user interface provided.
- No customized error messages.
Testing has been limited and users should expect potential errors. TEI import is still in development.
The following edited screenshot taken of the file doreco_teop1238_Gol_01.eaf from the DoReCo corpus version 2.0 for the language Teop in ELAN illustrates Corflow's model:
Corflow is built around a Transcription class used for universal information storage: all information from all the supported formats fit in. Import scripts/functions, e.g. fromElan, instantiate a Transcription object and fill it with the file's information; export scripts/functions, e.g. toElan, use a Transcription object to write a file. Manipulations are expected to operate on Transcription objects (after the import and before the export). In practice, this can vary as manipulations are open and dependent on the user's needs.
Generally, a transcription is for oral linguists text aligned to sound whereby the alignment relies on two time points. This notion of a transcription is captured in Corflow by the Segment class. A Segment object consists of text (content) with a start and end time. Segments might not be linguistic units, and might not be units at all (and conversely, a linguistic unit like the pause might have no corresponding Segment). A set of Segments corresponds to a Tier object and a set of Tiers corresponds to the whole Transcription. We don't claim here that all tiers, that is, all sets of segments, are linguistic transcriptions. They can also represent translations, annotations, etc. Tiers, like segments, are type-neutral.
Transcriptions, tiers and segments contain many more information and allow to access this information using different attributes and methods. For example, the metadata attribute contains all information around the transcription: where, when, who, how, ... The parent method and similar methods capture the hierarchical relations between segments and tiers. To learn more about the different attributes and methods available, visit the Github's wiki and take a look at the Corflow tutorial series provided by the AIRAL project.
The question of file conversion might never be answered in a satisfactory manner. Originally just an nth homemade conversion tool, our hope is Corflow becomes an easily accessible package for other teams and projects to use either as is, for basic use, or by being able to quickly adapt it to their requirements.
Corflow was created and is developed by François Delafontaine, and is actively developed and maintained by Aleksandr Schamberger.
DoReCo 2.0 database:
- Seifart, Frank, Ludger Paschen & Matthew Stave (eds.). 2024. Language Documentation Reference Corpus (DoReCo) 2.0. Lyon: Laboratoire Dynamique Du Langage (UMR5596, CNRS & Université Lyon 2). DOI:10.34847/nkl.7cbfq779
DoReCo 2.0 Teop dataset:
- Mosel, Ulrike. 2024. Teop DoReCo dataset. In Seifart, Frank, Ludger Paschen and Matthew Stave (eds.). Language Documentation Reference Corpus (DoReCo) 2.0. Lyon: Laboratoire Dynamique Du Langage (UMR5596, CNRS & Université Lyon 2). https://doreco.huma-num.fr/languages/teop1238 (Accessed on 14/02/2025). DOI:10.34847/nkl.9322sdf2
Methods used in building DoReCo:
- Paschen, Ludger, François Delafontaine, Christoph Draxler, Susanne Fuchs, Matthew Stave & Frank Seifart. 2020. Building a Time-Aligned Cross-Linguistic Reference Corpus from Language Documentation Data (DoReCo). In Proceedings of The 12th Language Resources and Evaluation Conference, 2657–2666. Marseille, France: European Language Resources Association. https://www.aclweb.org/anthology/2020.lrec-1.324 (2024/03/05).
Corflow and this repository are licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. For a quick review of the license, visit the license's website.